The unacceptably high rates of youth smoking, together with the failure to date of school-based interventions to effect long-term reductions in youth smoking, argue for new, innovative, theory-driven investigations of the youth smoking acquisition process. Responding to this need, this study will use a novel statistical model based on epidemic-type transmission of smoking, and high quality data resources from a well-maintained, 10-year cohort of 4,211 children, to investigate (1) to what extent smokers in each of six social environments of the child can predict youth smoking acquisition and progression through smoking stages, (2) to what extent certain child personality characteristics and home/school smoking policies act as effect modifiers for these predictions, and (3) to what extent selected tobacco-specific beliefs act as mediators. The study design employs a new and potentially productive epidemic model of youth smoking acquisition. The model is grounded in Social Cognitive Theory (SCT) [Bandura, 1986], positing that behavior is primarily derived from social exposures. In terms of infectious disease epidemics, infected persons (i.e., "smokers") in the child's social environment transmit the disease ("smoking behavior") to the uninfected ("non- smoking") child. The Investigations capitalize on availability of data on smoking in six social environments of the child: parents, siblings, close friends, same-grade classmates, older schoolmates, and school staff. The population-based nature of the cohort will yield generalizable results. [unreadable] This study design will -- quickly and efficiently -- yield critically needed and easily interpretable information on the simultaneous influences of smokers in the child's various social environments on both smoking and progressions through smoking stages, and on Interactions with personality and school/home smoking policies. The results can be expected to help contribute to the smoking and cancer mortality reduction goals of the National Cancer Program. [unreadable] [unreadable]